Autotune: fix inputs #926
Merged
Autotune: fix inputs #926
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #926 +/- ##
==========================================
- Coverage 86.43% 86.43% -0.01%
==========================================
Files 471 470 -1
Lines 44309 44326 +17
==========================================
+ Hits 38300 38314 +14
- Misses 6009 6012 +3
☔ View full report in Codecov by Sentry. |
nathanielsimard
requested changes
Nov 3, 2023
nathanielsimard
approved these changes
Nov 6, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Batch sizes
At first we thought batch sizes would have no incidence on kernels, i.e. the time taken for many batches would just be proportional to the time for one batch. So for autotune, to run faster, we had reduced the number of batches to a constant. However, our tests have shown that the matmul algorithms have different behaviours on different batch sizes, therefore it is important to keep them into the computations.
Random inputs
Before, we had only tensors of ones in autotune. It is possible that there will someday exist algorithms that have a biased behaviour on such trivial inputs, so it has to be replaced by random inputs. I chose uniform in [-10., 10.] but that's arbitrary.